Remember when they claimed Yi had 200k context length despite it having 16k of usable context? I remember, because I spent non-trivial effort trying to make it

Remember when they claimed Yi had 200k context length despite it having 16k of usable context? I remember, because I spent non-trivial effort trying to make it work for long-form technical summarization. My lackluster findings were validated by RULER. https://github.com/hsiehjackson/RULER https://github.com/hsiehjackson/RULER